Sep 26, 2025 · 1h 44m · bg2-pod
NVIDIA: OpenAI, Future of Compute, and the American Dream | BG2 w/ Bill Gurley and Brad Gerstner · Bg2 Pod
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the BG² podcast, host Brad Gerstner and partner Clark Tang interview NVIDIA CEO Jensen Huang at NVIDIA headquarters to explore the scaling laws of artificial intelligence, global compute infrastructure requirements, and NVIDIA's strategic moat. The conversation spans AI's macroeconomic impacts, US-China technological competition, sovereign AI initiatives, and policies aimed at expanding American economic prosperity.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Brad and Bill hold 32.6% of the talking time here. How this is scored →
Brad and Bill as informed peer 6.5 Guest teaching 4.9 Guest disagreement 2.0 Brad and Bill pushing back 2.6
speaking balance: gold is Brad and Bill, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:25:40 Badge of shame for China hawks
Jensen forcefully denounces the 'China hawk' label as a badge of shame, aggressively arguing that choking the talent pipeline and cutting off global competition undermines genuine patriotism.
Hardest push from Brad and Bill ▶ 8:41 Challenging growth with Wall Street consensusBrad directly confronts Jensen with consensus data from 25 sell-side analysts showing Nvidia's revenue growth flatlining to 8% by 2027, refusing to accept an unchallenged optimistic outlook.
Biggest teaching moment ▶ 51:11 The free-chip TCO mathJensen systematically educates the hosts on the financial physics of data centers, proving that when power is the bottleneck, inferior perf-per-watt makes competing chips cost-prohibitive even at zero purchase price.
Brad and Bill hold their own ▶ 27:37 Dissecting circular revenue and bubble fearsBrad demonstrates deep domain knowledge by citing financial precedents like Nortel/Cisco and asking Jensen to directly delineate genuine consumer demand from potential round-tripping.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Brad and Bill as informed peer | Guest teaching | Guest disagreement | Brad and Bill pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome to NVIDIA & Red Glasses Banter | 5 | 5 | 2 | 1 | Brad recaps their previous conversation regarding inference scaling and notes current revenue statistics. Jensen reframes the scaling narrative by introducing three distinct scaling laws (pre-training, post-training, and inference reasoning). | |
| NVIDIA's $100 Billion OpenAI Stargate Partnership | 6 | 4 | 1 | 2 | Brad asks detailed questions about the OpenAI Stargate partnership and potential four hundred billion dollar revenue implications. Jensen explains the structural rationale of partnering across Azure, OCI, CoreWeave, and direct self-building. | |
| Self-Building Hyperscalers and Direct Infrastructure Control | 7 | 3 | 2 | 5 | Brad challenges Jensen with Wall Street consensus estimates predicting Nvidia's growth will flatline to 8 percent by 2027. Jensen dismisses the skepticism, noting that Nvidia regularly beats consensus and builds for foundational opportunity. | |
| Three Pillars of Accelerated Computing & Macro GDP Augmentation | 6 | 6 | 1 | 1 | Jensen lays out a detailed macro thesis for why general purpose computing is obsolete and calculates how augmenting global GDP creates a multi-trillion dollar infrastructure TAM. Clark and Brad contribute TAM numbers and historical economic context. | |
| Bubble Fears, Compute Gluts, and Supply Chain Realities | 7 | 5 | 3 | 6 | Brad pushes hard on bubble fears, prisoner's dilemma dynamics, and potential compute gluts over the next 3 to 5 years. Jensen rejects the premise, asserting that the transition of all traditional hyperscaler workloads to AI means the market is already at massive scale. | |
| Accelerated Data Processing & Extreme Full-Stack Co-Design | 7 | 6 | 2 | 4 | Brad brings up media skepticism regarding circular revenues and round-tripping reminiscent of Cisco and Nortel. Jensen clarifies the financing mechanisms of compute buyers and transitions into the engineering feats of extreme full-stack co-design. | |
| NVIDIA's Competitive Moat and Ecosystem Strategy | 7 | 6 | 3 | 3 | Clark and Brad question Jensen on the viability of ASICs (like Google TPUs and AWS Trainium) versus GPUs. Jensen explains that custom ASICs cannot keep pace with rapidly changing models and the massive full-system AI factory scale. | |
| Total Cost of Ownership & Free Compute Economics | 7 | 6 | 2 | 3 | Brad asks Jensen to justify his provocative claim that even if competitor chips were free, customers would still buy Nvidia. Jensen explains total cost of ownership, showing power constraints make perf-per-watt opportunity costs overwhelmingly decisive. | |
| Long-Term Market Cap & AI Infrastructure Vision | 6 | 4 | 1 | 2 | Brad posits that Nvidia could become the world's first ten trillion dollar company and asks if revenue can double or triple over 5 years. Jensen emphasizes that Nvidia is an infrastructure partner rather than merely a component supplier. | |
| Elon Musk and the Speed of xAI's Colossus | 6 | 5 | 1 | 1 | Brad and Jensen discuss Elon Musk's speed in deploying xAI's Colossus cluster and the broader geopolitical imperative of sovereign AI infrastructure. | |
| Nvidia's China Strategy and AI Competition | 7 | 6 | 4 | 3 | Jensen strongly rejects prevailing political narratives that China is years behind in AI and technology, warning that export restrictions empower competitors like Huawei. Brad highlights the economic and technological trade-offs facing US policy. | |
| H-1B Visas, US Immigration Policy, and AI Talent Pipeline | 6 | 5 | 3 | 3 | Clark and Brad raise the proposed 100k H-1B visa fees and risks to the US talent pipeline. Jensen labels anti-immigration China hawks as a badge of shame and stresses that welcoming top global STEM talent is critical to maintaining American leadership. | |
| Pragmatic Diplomacy, US-China Relations, and Coexistence | 6 | 4 | 2 | 2 | Brad and Jensen discuss the difference between pragmatic dealmaking and dogmatic zero-sum geopolitical competition, using Nvidia's partnership with historical rival Intel as an analogy. | |
| The Invest America Act and Expanding Wealth Ownership | 7 | 4 | 2 | 2 | Brad discusses the Invest America Act passed in recent legislation to seed investment accounts for newborn children, which Nvidia supports. Jensen expands on re-industrialization, skilled craft labor, and why AI will create more ideas and jobs rather than causing net job destruction. | |
| Exponential Future, Robotics, and Responsible Leadership | 7 | 5 | 1 | 1 | Brad and Jensen close on exponential technological compounding, the fusion of AI with robotics, personalized digital twins, and the duty of industry leaders to bring the broader public along. |
Statements from this episode (44)
Gerstner: Over 40% of NVIDIA's revenue comes from AI inference workloads
“Over 40% of your revenue today is inference.”
Huang: Inference Compute Demand Will Surge One Billion Times
“It's about to go up by a billion times.”
Huang: AI Now Governed by Three Scaling Laws, Not One
“And so thinking, post-training, pre-training, we now have three scaling laws, not one.”
Huang: OpenAI is likely the next multi-trillion dollar hyperscaler
“I think that OpenAI is likely going to be the next multi-trillion dollar hyperscale company. Hyperscale like, ah, like Meta's a hyperscale. Google's a hyperscale. They're going to have consumer and enterprise services, and, ah, they are very likely going to be…”
Huang: OCI, OpenAI, and SoftBank have contracted 5 to 7 gigawatts
“The second is the OCI build out. And I think there's some five, six, seven gigawatts about to be built out. And so working with OCI and OpenAI and SoftBank to build that out. Those projects are contracted.”
Huang: NVIDIA is partnering directly with OpenAI to self-build infrastructure
“This new partnership is about helping OpenAI, working, partnering with OpenAI to build their own self-build AI infrastructure for the first time. And so this is us working directly with OpenAI at the chip level, at the software level, at the systems level, at …”
Huang: OpenAI and xAI will likely consume compute capacity internally
“I think they'll likely use it themselves and just in the case of X, they'll likely use it themselves.”
Huang: Hyperscalers support NVIDIA's direct purchasing deals with AI labs
“But they would like to have the same direct relationship with us, direct working relationship, and direct purchasing relationship. Meta, just as what Zuck and Meta has with us it's exactly a direct. Our relationship with between us and Sundar and Google, direc…”
Gerstner: Wall Street consensus projects NVIDIA growth dropping to 8% in 2027
“But of the 25 sell-side analysts on Wall Street who cover your stock, if I look at the consensus estimate, it basically has your growth flatlining starting in twenty-twenty-seven. Eight percent growth twenty-twenty-seven through twenty-thirty, ok?”
Huang: NVIDIA is comfortable with low Wall Street consensus estimates
“Look, we're comfortable with that. Beating the numbers on a regular basis.”
Gerstner: Massive divergence remains between tech CEOs and Wall Street on AI
“And what it shows me is still, even though we're two and a half years into the age of AI, a massive divergence of belief between what we hear Sam Altman saying, you saying, Sundar saying, Satya saying, and what Wall Street still believes.”
Huang: General-purpose computing is over, replaced by accelerated AI computing
“That general, general purpose computing is over, and the future is accelerated computing and AI computing.”
Huang: 100% of NVIDIA engineers and chip designers use AI agents
“Every single software engineer, every single chip designer in our company already has AIs working with them. A hundred percent coverage.”
Huang: $5 trillion annual global AI infrastructure CapEx is economically plausible
“So if you told me that on an annual basis, the CapEx of the world was about five trillion dollars, I would say the math seems to make sense.”
Huang: Alibaba CEO Eddie Wu plans to 10x data center power by 2030
“Eddie, last night, Eddie Wu at Alibaba said, between now and the end of the year, and excuse me, now and the end of the decade, they're going to increase their data center power by 10 X, right?”
Huang: AI compute glut unlikely until general-purpose computing fully transitions
“Until, until we fully Convert all general purpose computing to accelerated computing in AI until we do that. I think the chances are extremely low.”
Huang: Hyperscalers have consistently underforecasted GPU demand for years
“Now what's, what's going on, of course, is that every one of their forecasts that they provide us, it turns out to have been wrong. Because they under forecasted, and so now we're always in a scramble mode, and so we've been in the scramble mode now for, you k…”
Huang: Data processing at Snowflake and Databricks runs mostly on CPUs
“Data processing represents the vast majority of the world's CPUs today. It still completely runs on CPUs. You know, if you go to Databricks, it's mostly CPUs. If you go to Snowflakes, mostly CPUs. SQL processing at Oracle, mostly CPUs. Everybody's using CPUs t…”
Huang: Chipmakers must co-design six to eight chips annually to beat wafer costs
“Wafer costs are getting higher, which means that unless you do co-design at an extreme scale, you're just not going to be able to deliver the x-factor growth, number one, number, and so you know, unless you, unless you're working on six, seven, eight chips a y…”
Huang: A One-Gigawatt AI Data Center Requires 400,000 to 500,000 GPUs
“When your customers deploy a gigawatt, that's a, you know, 400,500 thousand GPUs. Getting 500,000 GPUs to work together is a miracle.”
Huang: Startups should capture entire nascent markets rather than entering mature ones
“You're supposed to create a startup before the market grows. You're not supposed to come up as a startup when the market's a trillion dollars large. You know, this fallacy, and all VCs know this fallacy that a large market, if you could just take a few percent…”
Huang: Google TPUs will inevitably transition to customer-owned tooling at scale
“Where will TPUs go when they, when it becomes a large business? Customer-owned tooling. There's no question about it.”
Huang: Operators will reject free AI chips to maximize performance per watt
“So even if they gave it to you for free, you only have two gigawatts to work with, your opportunity cost is so insanely high, you would always choose the best perf per watt.”
Gerstner: NVIDIA will likely be the first $10 trillion company
“I think NVIDIA will likely be the first 10 trillion dollar company”
Huang: NVIDIA does not force customers into bundled full-rack systems
“We don't require anybody to buy everything from us. We don't require that they buy the full rack. They could buy a chip. They could buy a component. They could buy our networking. They could buy our, we have customers buying only our CPU. You know, just buy ou…”
Huang: Elon Musk will likely build a gigawatt AI cluster first
“I would not be surprised if he gets to a gigawatt before anybody else does.”
Huang: AI supercomputers are humanity's most complex systems problem ever
“I mean, these are, this is unquestionably the most complex systems problem humanity has ever endeavored.”
Huang: Every country must develop sovereign AI capabilities
“And so I think that, that every country needs to have some sovereign capability. I recommend that they all use OpenAI, they all use Gemini, they all use, you know, these, Open models. You use Grok, and I think they, I recommend they all do that. I recommend th…”
Huang: President Trump is right to maximize US tech exports
“I think President Trump's got it right, that we want to maximize exports. We want to maximize American influence around the world. We're supposed to maximize those things.”
Huang: Commerce Secretary Howard Lutnick is actively accelerating US tech exports
“Secretary Lutnik was all over it.”
Huang: China is 'nanoseconds behind' the US in AI capability
“And three, they're years behind us. Is it two years, three years? Come on, they're nanoseconds behind us.”
Huang: Nvidia's financial guidance assumes zero revenue from China
“Even though I tell all of our investors. That our guidance includes no China. And I appreciate all of our investors to include no China in any of our guidance.”
Huang: Nvidia will eventually get the opportunity to compete in China
“These things will get sorted out, and we will have the opportunity to go compete in that China market.”
Huang: $100k H-1B fee might accelerate investment outside the US
“It also has an unintended consequence. It might accelerate investment outside United States.”
Huang: Drop in Chinese students staying in US is existential crisis
“Definitely see a greater concern of Chinese students who come here and remain here. And, or many of them who come here for school and are thinking about going elsewhere. Many of them thinking about Europe. And so, so I think we need to be super, super concerne…”
Huang: Restricting US immigration as a China hawk is unpatriotic
“It's almost like a badge of honor. It's a badge of shame. There's no question it's a badge of shame. There's no question that although they want what's in the best interest of our country, and we all want what's in the best interest of our country, destroying …”
Gerstner: Donald Trump will reach a big deal with China
“The reason I'm confident and I've said on this pod, That I think he'll get a big deal done with China.”
Huang: Decoupling the US and China is entirely the wrong concept
“You can't decouple against the single most The two most important relationships for the next century. That doesn't make any sense at all. Decoupling is exactly the wrong concept.”
Huang: NVIDIA partnered with Intel despite historic rivalry
“And notice, we just did the ultimate deal. Right. Partnering with Intel. A company that spent most of its life trying to put us out of business. And I had no trouble partnering with them.”
Gerstner: Tech will produce $10T companies and probably trillionaire individuals
“Because of the nature of these technological systems, we have companies that are going to be worth 10 trillion. We'll probably have individuals that are worth a trillion.”
Huang: NVIDIA is hiring across the board as AI boosts productivity
“My engineers, every one of them use AIs continuously, and we're hiring more engineers, we're hiring more people, we're hiring across the board. The reason for that is because we have more ideas. We can now go pursue more ideas. The reason for that is because o…”